• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.150-161
• /
• 2018

We propose a novel deep neural network model for detecting human activities in untrimmed videos. The process of human activity detection in a video involves two steps: a step to extract features that are effective in recognizing human activities in a long untrimmed video, followed by a step to detect human activities from those extracted features. To extract the rich features from video segments that could express unique patterns for each activity, we employ two different convolutional neural network models, C3D and I-ResNet. For detecting human activities from the sequence of extracted feature vectors, we use BLSTM, a bi-directional recurrent neural network model. By conducting experiments with ActivityNet 200, a large-scale benchmark dataset, we show the high performance of the proposed DeepAct model.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.637-640
• /
• 2023

This paper presents multimodal image fusion with human pose for detecting abnormal human behaviors in low illumination conditions. Detecting human behaviors in low illumination conditions is challenging due to its limited visibility of the objects of interest in the scene. Multimodal image fusion simultaneously combines visual information in the visible spectrum and thermal radiation information in the long-wave infrared spectrum. We propose an abnormal event detection scheme based on the multimodal fused image and the human poses using the keypoints to characterize the action of the human body. Our method assumes that human behaviors are well correlated to body keypoints such as shoulders, elbows, wrists, hips. In detail, we extracted the human keypoint coordinates from human targets in multimodal fused videos. The coordinate values are used as inputs to train a multilayer perceptron network to classify human behaviors as normal or abnormal. Our experiment demonstrates a significant result on multimodal imaging dataset. The proposed model can capture the complex distribution pattern for both normal and abnormal behaviors.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2014.02a
• /
• pp.429.1-429.1
• /
• 2014

We present a method of surface functionalization of a single layer graphene for linking and detecting MDA-MB-231 human breast cancer cell. The methodology is done by utilizing 1-pyrenebutanoic acid and succinimidyl ester for immobiling CD44 antibodies. This work shows that the single layer graphene is an efficient fixing substance to capture the MDA-MB-231 human breast cancer cell, selectively. The immobilization method of the cancer cell on the graphene layer will be an effective cell counting system. Moreover usage of the linking with non-covalent bonding is expected to develope a sensor scheme of electrical cell-detecting diagnosis system.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• v.29 no.1
• /
• pp.493-498
• /
• 2005

In this paper, we deal with on observer design for detecting the human faults in container crane operation. First we propose an observer for detecting the human faults and show the existing condition for the observer. In this case, we assume that the human faults can be considered ad a careless mistake during the crane operation. In simulation, we used the previous results for human work model and design the observer for the human work model. As a simulation results with human faults, the proposed observer can detected the human faults perfectly, thus the efficiency of proposed observer is shown.

• International Journal of Contents
• /
• v.2 no.1
• /
• pp.5-8
• /
• 2006

This paper describes a new mechanism of detecting object of interest from a noisy image, without using any a-priori knowledge about the target. It employs a parallel set of filters inspired upon biological findings of mammalian vision. In our proposed system, several basic features are extracted directly from original input visual stimuli, and these features are integrated based on their local competitive relations and statistical information. Through integration process, unnecessary features for detecting the target are spontaneously decreased, while useful features are enhanced. Experiments have been performed on a set of computer generated and real images corrupted with noise.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.24 no.12
• /
• pp.977-984
• /
• 2014

Nanomaterial have grown from scientific interest to commercial products and the nanomaterial market has grown 19.1 % each year. As the nanomaterial market size increases, it is expected that nanomaterial production will increase and its contamination of outdoor environmental system will also increase in the form of industrial waste. Since most of nanomaterials are known as biologically non-degradable materials, nanomaterials will accumulate in the environment, and this will increase the potential threats to human health along the food chain. Recent studies have investigated the toxicity effect of nanomaterials due to their size, chemical composition and shape. For the development of nanomaterial while taking human health into consideration, a nanomaterial detecting sensor is required. In this paper, we have observed the trend of nanomaterial detecting sensor of mechanical, electrochemical, optical and kelvin probe force microscopy sensors and we believe that this trend will shed the light on the development of real-life nanomaterial detecting sensors.

• International Journal of Internet, Broadcasting and Communication
• /
• v.14 no.4
• /
• pp.88-95
• /
• 2022

Detecting anomalies in human movement is an important task in industrial applications, such as monitoring industrial disasters or accidents and recognizing unauthorized factory intruders. In this paper, we propose an abnormal worker movement detection system based on data stream processing and hierarchical clustering. In the proposed system, Apache Spark is used for streaming the location data of people. A hierarchical clustering-based anomalous trajectory detection algorithm is designed for detecting anomalies in human movement. The algorithm is integrated into Apache Spark for detecting anomalies from location data. Specifically, the location information is streamed to Apache Spark using the message queuing telemetry transport protocol. Then, Apache Spark processes and stores location data in a data frame. When there is a request from a client, the processed data in the data frame is taken and put into the proposed algorithm for detecting anomalies. A real mobility trace of people is used to evaluate the proposed system. The obtained results show that the system has high performance and can be used for a wide range of industrial applications.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2005.06a
• /
• pp.1131-1135
• /
• 2005

Since human is vulnerable to emotional, physical and environmental distractions, most human inspectors cannot sustain a consistent 8-hour inspection in a day specifically for small components like door locking levers. As an alternative for human inspection, presented in this study is the development of a machine vision inspection system (MVIS) purposely for door locking levers. Comprises the development is the structure of the MVIS components, designed to meet the demands, features and specifications of door locking lever manufacturing companies in increasing their production throughput upon keeping the quality assured. This computer-based MVIS is designed to perform quality measures of detecting missing portions and defects like burr on every door locking lever. NI Vision Builder software for Automatic Inspection (AI) is found to be the optimum solution in configuring the needed quality measures. The proposed software has measurement techniques such as edge detecting and pattern-matching which are capable of gauging, detecting missing portion and checking alignment. Furthermore, this study exemplifies the incorporation of the optimized NI Builder inspection environment to the pre-inspection and post-inspection subsystems.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2001.04a
• /
• pp.243-248
• /
• 2001

A radial artery pulse wave is well known as a good mans to diagnose human body condition in th field of Chinese medical science. Information about constitution as well as organs can be obtained by detecting the artery pulse wave. Recently, the information about the human body constitution may be utilized in accelerating the recovery process of the patient on the basis of comprehensive diagnosis. A radial artery pulse wave is considered as one of promising means in examining the human body constitution. Since the examination has been conducted by the feeling of finger, the diagnosis may occasionally have uncertainty or fatal error. In this paper, a new measuring system is suggested and developed to examine the pattern of a pulse wave correctly. The system is composed of four pressure vessels, pressure sensors and air supplying pumps. One of them is utilized for appropriately pressing the radial artery, three of them for detecting pressure change in several mmHg level. The detected data is shown and discussed.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2319-2322
• /
• 2003

We propose ROI(region of interest) image coding application to still image using PSBS(partial significant bitplane shift)method combined with human face region detecting system. PSBS is an encoding algorithm for ROI image coding in JPEG2000, and takes advantages of both generic scaling based method and maximum shift method defined in JPEG2000. The Powerful advantages of PSBS are able to adjusting image quality in ROI and background flexibly, and support arbitrarily shaped ROI coding without coding the shape. In this letter, we show how to compress an image for human face region using PSBS method combined with human face region detecting system, and propose its application.